Essential Points
- Qualcomm and Siemens demonstrated a live autonomous factory model at MWC Barcelona 2026, combining edge AI with private Industrial 5G
- AGVs and robotic arms coordinated in real time over a Siemens private 5G network powered by Qualcomm technology
- Qualcomm Cloud AI 100 Accelerator Card ran worker assistance, diagnostics, quality inspection, and workflow optimization entirely on-premises
- Qualcomm’s Dragonwing Q-8750 delivers up to 77 TOPS of AI performance and supports on-device LLMs up to 11 billion parameters
The factory floor just changed. Qualcomm and Siemens jointly demonstrated a fully operational autonomous production model at MWC Barcelona 2026, showing how private 5G networks and on-premises AI can execute real-time manufacturing decisions without cloud dependency. This is not a prototype concept but a live showcase of technology deployable today, built on hardware already shipping to industrial buyers.
What Qualcomm and Siemens Demonstrated at MWC Barcelona 2026
The demonstration brought together two technologies: Siemens’ private Industrial 5G network and Qualcomm’s edge AI compute stack. An Automated Guided Vehicle transported materials across the factory model while communicating with robotic arms and production cells, all coordinated over a low-latency 5G link.
A Siemens industrial PC equipped with the Qualcomm Cloud AI 100 Accelerator Card ran an AI agent locally, analyzing system state in real time to identify root causes of production issues and recommend corrective actions. That same card handled real-time process quality inspection and on-premises diagnostics.
The combination keeps production data entirely within factory premises, addressing security and compliance requirements that cloud-dependent deployments cannot satisfy.
The Qualcomm Cloud AI 100 Card: Four Industrial Functions in One Device
The Qualcomm Cloud AI 100 Accelerator Card is the hardware anchor of this factory AI deployment. Installed inside a standard Siemens industrial PC, it delivers four AI-driven functions confirmed in the MWC demonstration: worker assistance, diagnostics, quality inspection, and workflow optimization.
It operates as a drop-in upgrade to existing industrial PC infrastructure rather than requiring a full system replacement. This means manufacturers can add on-premises AI capabilities to current production environments without rebuilding their hardware stack from the ground up.
Nakul Duggal, EVP and Group General Manager of Automotive, Industrial, Embedded IoT, and Robotics at Qualcomm Technologies, stated: “By combining Qualcomm Technologies’ leadership in AI, edge computing and connectivity with Siemens’ expertise in factory automation, we are showcasing a new class of intelligent, autonomous manufacturing powered at the edge.”
Private 5G as the Industrial Network Layer
The Siemens Industrial 5G network in the demonstration is a private network engineered for mission-critical factory operations. Qualcomm has powered Siemens’ private 5G infrastructure across multiple industrial deployments as part of a multi-year partnership focused on flexible, mission-critical wireless connectivity.
The combination of this private network and on-premises AI allows manufacturers to automate production, improve quality, and keep sensitive production data within their own premises. Thibault de Assi, SVP at Siemens AG, confirmed the joint direction: “This showcase highlights a shared vision for the next major productivity leap in industry, where real-time intelligence drives efficiency, resilience, and long-term value for manufacturers.”
Qualcomm Dragonwing Q-8750: The Processor Behind Industrial Edge AI
The Dragonwing Q-8750 is Qualcomm’s most advanced IoT processor as of early 2026. It delivers up to 77 TOPS of AI performance and supports real-time inference with on-device large language models up to 11 billion parameters, reducing reliance on cloud connectivity for critical applications.
Its camera architecture supports up to 12 physical cameras simultaneously, making it suited for drones, media hubs, and complex multi-point vision systems in production environments. The processor is designed to run critical workloads locally, including applications in regulated industries where cloud connectivity is restricted or prohibited.
Edge Impulse, acquired by Qualcomm over the past 18 months, is now fully integrated into the Qualcomm Dragonwing AI On-Prem Appliance, enabling organizations to train and deploy large AI models in offline or private-network environments.
Qualcomm’s Broader Industrial Build-Out in 2026
At CES 2026, Qualcomm announced five acquisitions completed over the previous 18 months: Augentix, Arduino, Edge Impulse, Focus.AI, and Foundries.io. Together these additions give Qualcomm developer tooling that spans robotics prototyping, vision AI, on-device model training, embedded OS support, and AIoT application deployment.
Qualcomm’s goal, as stated by Nakul Duggal at CES 2026, is to help organizations of all sizes deploy “scalable, intelligent systems at the edge” through a unified software architecture supporting Linux, Windows, and Android.
At Embedded World 2026 (March 10-12, Nuremberg), SECO is demonstrating edge AI solutions built on Qualcomm Dragonwing platforms including the SOM-COMe-CT6-Dragonwing-IQ-X. This COM Express Type 6 module is powered by 12- or 8-core Qualcomm Oryon CPUs with up to 3.4 GHz single-thread performance and up to 45 TOPS of on-device AI acceleration. It supports real-time control, analytics, and high-resolution displays up to 2x5K or 4x4K, and operates in harsh environments from -40°C to +85°C.
SECO’s roadmap includes SMARC solutions based on Qualcomm Dragonwing IQ8 and COM-HPC Mini platforms based on Qualcomm Dragonwing IQ9, both planned for availability by Q3 2026.
At the network infrastructure level, Qualcomm launched the X105 5G Modem-RF at MWC 2026, the world’s first 3GPP Release 19-ready modem, laying groundwork for future 6G development.
Considerations
On-premises AI deployments require upfront capital investment in accelerator hardware, private 5G network infrastructure, and IT personnel with edge AI expertise. For smaller manufacturers, this creates a real cost barrier compared to pay-per-use cloud models. As of March 2026, Qualcomm and Siemens have not published benchmarked cost-per-deployment figures for the showcased stack, making independent ROI calculation difficult at this stage.
Qualcomm’s Agentic RAN Management Service: The Architecture Shift Telcos Can No Longer Ignore
Frequently Asked Questions (FAQs)
What did Qualcomm and Siemens demonstrate at MWC Barcelona 2026?
They demonstrated a live autonomous factory model featuring AGVs and robotic arms coordinating over a Siemens private Industrial 5G network. An on-premises AI agent running on the Qualcomm Cloud AI 100 Accelerator Card handled real-time diagnostics, quality inspection, worker assistance, and workflow optimization.
What is the Qualcomm Cloud AI 100 Accelerator Card?
It is an AI accelerator installed inside a Siemens industrial PC that delivers four on-premises functions: worker assistance, diagnostics, quality inspection, and workflow optimization. It runs entirely locally without cloud dependency, keeping production data within factory premises.
What are the confirmed specs of the Qualcomm Dragonwing Q-8750?
The Dragonwing Q-8750 delivers up to 77 TOPS of AI performance and supports on-device large language models up to 11 billion parameters. It supports up to 12 physical cameras simultaneously and is designed for drones, media hubs, and complex industrial vision systems.
What acquisitions did Qualcomm complete to support its industrial AI strategy?
Qualcomm completed five acquisitions over 18 months leading into CES 2026: Augentix, Arduino, Edge Impulse, Focus.AI, and Foundries.io. These add vision AI, robotics prototyping, on-device model training, AIoT OS support, and embedded developer tools to the Dragonwing platform.
What is SECO demonstrating at Embedded World 2026?
SECO is showcasing Qualcomm Dragonwing-based edge AI solutions at Embedded World 2026 in Nuremberg (March 10-12). Key products include the SOM-COMe-CT6-Dragonwing-IQ-X with up to 45 TOPS and Oryon CPUs running to 3.4 GHz, alongside MUSAI, a modular underwater AI system for real-time edge inference.
When will SECO’s SMARC IQ8 and COM-HPC Mini IQ9 solutions ship?
SECO confirmed both the SMARC IQ8 and COM-HPC Mini IQ9 solutions are planned for availability by Q3 2026. These platforms are designed for advanced industrial automation, next-generation HMIs, and Clea-enabled edge AI applications.
What is the Qualcomm X105 5G Modem announced at MWC 2026?
The Qualcomm X105 5G Modem-RF is the world’s first 3GPP Release 19-ready modem, announced at MWC Barcelona 2026. It lays the technical foundation for eventual 6G development while delivering advances in 5G Advanced performance for industrial and consumer applications.
What is the role of Edge Impulse in Qualcomm’s industrial AI stack?
Edge Impulse, acquired by Qualcomm over the past 18 months, is now integrated into the Qualcomm Dragonwing AI On-Prem Appliance. It enables organizations to train and deploy large AI models in fully offline or private-network environments, addressing requirements in critical infrastructure and regulated industries.
